Fish embryo and juvenile size under hypoxia in the mouth...

Current Zoology 58 (3): 401−412, 2012

Received May 9, 2011; accepted Oct. 26, 2011.

∗ Corresponding author. E-mail: [email protected]. Current Address: Biosciences, College of Life and Environmental Sciences, Univer-sity of Exeter, Exeter, Devon, EX4 4QD, UK; Fax: 44 (0)1392-263700.

© 2012 Current Zoology

Fish embryo and juvenile size under hypoxia in the mouth- brooding African cichlid Pseudocrenilabrus multicolor

E.E. REARDON1*, L.J. CHAPMAN1,2 1 Department of Biology, McGill University, 1205 Dr. Penfield Ave, Montreal, Quebec, H3A 1B1, Canada 2 Wildlife Conservation Society, 2300 Southern Boulevard, Bronx, NY 10460, USA

Abstract We used a field survey and a laboratory rearing experiment to (a) examine response (size and survival) to life-long hypoxia in offspring of the African maternal mouth-brooding cichlid Pseudocrenilabrus multicolor victoriae (Seegers) and (b) explore the degree to which developmental response can be environmentally-induced. Embryo size metrics were quantified in 9 field populations across a range of dissolved oxygen (DO) concentrations. In the laboratory, first generation (F1) broods of low-DO origin were reared under high or low DO. Brooding period was quantified for the mothers; and egg size, egg metabolic rate and juvenile size-at-release were quantified in their (F2) offspring. The F2 offspring were split and grown for 3 months post-release under high or low DO, and juvenile size and survival were quantified. In the field survey, across stages, embryos from low-DO field populations were shorter and weighed less than embryos from high-DO populations. In the laboratory experi-ment, F2 eggs and juveniles-at-release from mother’s mouth did not differ in mass, length, survival regardless of development DO environment. However, juveniles diverged in size after leaving mother’s mouth, exhibiting smaller size when grown under low DO. Size differences in embryo size across field populations and divergence in embryo size after release from the mother’s mouth support predictions for smaller body size under hypoxia. There was no evidence for negative effects on survival of juveniles after 3 months. Brooding period was 16% shorter in females reared under low DO suggesting that hypoxia may accelerate embryo de-velopment. This work provides insights into how bearer fishes respond to hypoxic stress relative to fishes with no post-spawning parental care; a shorter brooding interval and smaller body size may provide an optimal solution to parent and embryo survival under hypoxia in brooding fishes [Current Zoology 58 (3): 401−412, 2012].

Keywords African cichlid, Lake Victoria, Life-history, Parental care, Dissolved oxygen

Hypoxia, or low dissolved oxygen (DO), can operate as a significant environmental stressor for water- breathing organisms (Pollock et al., 2007; Wu, 2009) given the high costs of acquiring oxygen from water relative to air (Lindsey, 1978; Wootton, 1990). Aquatic hypoxia occurs naturally, but it is becoming widespread as a consequence of eutrophication and pollution of wa-ter bodies and is now recognized as a large-scale threat to fresh and coastal waters. The increasing frequency of hypoxic events has resulted in changes in biodiversity, population decline, and production of extensive “dead zones” that affect commercially and recreationally im-portant fisheries (Pollock et al., 2007; Diaz and Breit-burg, 2009; Wu, 2009). These negative impacts of hy-poxia on wild fish populations have been associated with reproductive and developmental impairment (e.g. Wu, 2009). Several aspects of reproduction in fishes and invertebrates have shown impairment due to hypoxia, including gametogenesis, the number and quality of

gametes, reproductive behavior, fertilization success, impaired respiration, hatching success, and survivorship of the young (e.g. Garside, 1959; Hamor and Garside, 1976; Rombough, 1988; 2007; reviewed in Wu, 2009 and Mueller et al., 2011). However, the majority of the evidence for the impacts of hypoxia on reproduction and early-life stage development comes from short-term (hours to weeks) laboratory experiments. We understand far less about life-long effects of hypoxia on reproduc-tion and early life-history traits.

Several reviews of fishes and early life-history traits suggest that large egg/embryo size is positively associ-ated with fitness correlates such as higher survival, faster growth, and enhanced swim performance (Reznick, 1991; Heath and Blouw, 1998). However, much of the early life-history and maternal effects lite-rature for fishes supports the idea that eggs, embryos, and/or juveniles should actually be smaller under low DO, because they will have higher survival, and thus

402 Current Zoology Vol. 58 No. 3

higher fitness, than larger individuals at the same stage (e.g., Sargent et al., 1987; Hendry et al., 2001). The benefits of small size under hypoxia can be explained physiologically in terms of the surface area to volume ratio; where surface area limits the amount of oxygen that can be taken up, and volume (theoretically) repre-sents the oxygen demand (Krogh, 1959; Pauly, 1981). At small sizes the surface area to volume ratio is higher, which means that there is a larger surface area to meet the metabolic demand. In fish eggs, oxygen uptake oc-curs across the egg membrane. In hatched embryos and small juvenile fish, oxygen uptake occurs primarily across the skin, while the gills at early life stages func-tion predominately for ion exchange (Moyle and Cech, 1996). Thus, optimal egg and juvenile size under hy-poxia suggest a negative size-fitness function i.e., under hypoxia, smaller should be better in terms of fitness. However, there is some argument as to whether the whole egg is metabolically active (e.g., Einum et al., 2002; Kolm and Anesjo, 2005; Rombough, 2007), which could alter the above prediction.

In addition to the physiological advantages of small egg size under hypoxia, it is possible that small size facilitates faster development. While widely supported in the theoretical literature, empirical support for this prediction is contradictory (Marshall and Bolton, 2007). Several empirical studies of marine invertebrates, ma-rine fish, and amphibians do support the longer deve-lopment time in large eggs (Pauly and Pullin, 1988; Bradford, 1990 Marshall and Keough, 2008). However, others provide support for the opposite or no relation-ship between egg size and development within and across taxa (Marshall and Keough, 2008; Przeslawski and Webb, 2009; Ruis et al., 2010). Clearly, there is a need for additional empirical work that explores the effects of environmental stressors on egg size and de-velopment.

In this study, we used a field survey and a laboratory rearing experiment to (a) examine response (size and survival) to life-long hypoxia in embryos and juveniles of the African maternal mouth-brooding cichlid Pseu-docrenilabrus multicolor victoriae (Seegers) and (b) explore the degree to which developmental response can be environmentally-induced. Much of the work-to-date has been done predominately on batch spawning fishes and nest guarders (reviewed in Wu, 2009 and Mueller et al., 2011). However, bearer species (those that carry their offspring externally or internally) may have evolved different strategies to mitigate costs associated with reproduction and/or development under hypoxia

given that they carry their offspring throughout embry-onic development. P. multicolor was used to explore these questions because it is found across a wide range of oxygen habitats throughout the Lake Victoria basin of East Africa and exhibits strong interdemic variation in morphology, physiology, and life-history traits (e.g., Chapman et al., 2000; Chapman et al., 2002a,b; Chap-man et al., 2008; Martinez et al., 2009; Reardon and Chapman, 2009). In P. multicolor, it is unclear whether fertilization occurs on the substrate, in the mouth, or some combination of both (Wickler, 1962; Fryer and Iles, 1972; Hellweg, 2008). As with other bearer cichlid species, embryos undergo direct development (no larval stage; Balon, 1986; Noakes, 1991). Offspring are re-leased from the mouth as juveniles, when their yolk sacs are completely absorbed and they are ready to feed exogenously (Noakes, 1991).

We compared embryo size of 4 developmental stages of P. multicolor among 9 field populations. These popu-lations were selected from sites that differed in aquatic oxygen availability. In the laboratory, we reared first generation (F1) broods of low-DO origin under either high or low DO. Once the F1 generation reproduced, we quantified the brooding period (i.e. the length of time the mother held the young in her mouth from fertiliza-tion to release) and the F2 egg size, egg metabolic rate, and juvenile size at release (see Fig. 1 for experimental design schematic). The second generation (F2) juveniles from the laboratory experiment were then split and grown for 3 months post-release under either high or low DO. We predicted the following:

• Across field populations, embryo size will be positively correlated with DO.

• Egg size and juvenile size at release will be smaller when mothers are reared and embryos brooded (embryo development environment) un-der low DO compared to high DO.

• If egg and embryo size are smaller under low DO, then brooding period will be reduced because smaller embryos should develop faster. Juvenile size after 3 months of post-release growth will be smaller in juveniles grown under low DO com-pared to high DO.

1 Methods and Materials 1.1 Field survey

We collected brooding P. multicolor females in Uganda during the May-June dry season (2004–2007). Fish were sampled from 9 field populations across a range of DO concentrations (range: 0.24 to 8.7 mg L-1).

REARDON EE, CHAPMAN LJ: Fish embryo and juvenile size under hypoxia 403

Fig. 1 Schematic of the experimental design for development and growth environments of Pseudocrenilabrus multicolor in the laboratory rearing experiment First generation (F1) broods of low-DO (Lwamunda Swamp) origin were reared under either high or low DO. Once the F1 generation reproduced, we quantified the brooding period and the F2 egg size, egg metabolic rate, and juvenile size. The second generation (F2) juveniles from the labora-tory experiment were then split and grown for 3 months post-release under either high or low DO after which fry size and survival were measured.

We chose sites representative of the range of DO condi-tions experienced by this species (see Fig. 2 map, adapted from Reardon and Chapman, 2009 and Crispo and Chapman, 2010). These included: Lake Saka and four sites in the Mpanga River drainage of western Uganda, and four sites in the Nabugabo region of the Lake Victoria basin. The Mpanga River is located >150 km west of the Nabugabo system and joins the west-ward flowing section of the Katonga River feeding into Lake George. Environmental data were collected from multiple micro-sites within each location in the after-noon on days when fish were collected. Environmental variables included: DO, water temperature, conductivity, pH, and transparency (Table 1; see Reardon and Chap-man 2009 for full description). Aquatic dissolved oxy-gen concentration was converted from mg L-1 to partial pressure (mmHg), correcting for temperature and alti-tudinal differences across sites. These readings were then averaged to produce a mean DO level for each field population. Although P. multicolor breed throughout the year, there are peaks in reproductive activity associated with peaks in rainfall (Welcomme, 1969; Reardon and Chapman, 2008). Expeditions were carried out in the late May - early June period to standardize for potential seasonal effects on reproductive traits. We live-captured an average of 48 females per population per year sam-pled (range: 22-74) using metal minnow traps as part of a larger study estimating size at maturity and other

Table 1 Mean dissolved oxygen (DO) values for each of the field sampling sites where Pseudocrenilabrus multicolor were sampled (adapted from Reardon and Chapman, 2009)

Field Population Region Mean DO (mg L-1)

Bunoga Mpanga 7.32

Kahunge Mpanga 4.20

Kamwenge Mpanga 6.92

Kantembwe Mpanga 0.24

Laka Saka Mpanga 8.73

Lake Kayanja Nabugabo 6.30

Kaz Lagoon Nabugabo 1.63

Lwamunda Swamp Nabugabo 0.95

Lake Nabugabo Nabugabo 2.76

life-history traits in this species (reported in Reardon and Chapman, 2009). Brooding females and their broods were used to quantify egg traits (Reardon and Chapman, 2009) and size and number of developing embryos (this study). Among the females collected, offspring were in all stages of pre-release development (unhatched eggs to fully developed juveniles). However, females are very secretive while brooding, and we were not always able to capture all stages of development at every site. Thus, embryo size data were combined across years to increase power. We were unable to test for an effect of year on brood traits; but as stated above,


we controlled for season sampled to minimize seasonal effects. When a brooding female was live-captured, the brood was gently removed from the mouth. Each brood was counted for offspring number, euthanized in tricaine methanosulfate (MS222), and preserved in 4% parafor-maldehyde buffered with PBS. Egg size and number, female size at maturity, and environmental data were previously reported in Reardon and Chapman (2009).

Total length and mass were measured for each em-bryo and macroscopically classified into one of the fol-lowing five developmental stages: Stage 1 is the fertil-ized unhatched egg (reported in Reardon and Chapman, 2009); Stage 2 is the embryo hatched out of the egg with a very large yolk sac, on a small translucent body, with a small eye; Stage 3 is similar to Stage 2 with a reduced yolk sac, a larger eye but still a relatively small body; Stage 4, the yolk sac is almost completely absorbed; the eye fills a larger portion of the head; Stage 5, the yolk sac is completely gone (see Fig. 4). Between 474 and 1,009 individual hatched embryos were measured for each these five developmental stages. Two analytical approaches were used to detect relationships between DO and embryo size. First, analysis of covariance (ANCOVA) was used to test for an effect of develop-ment stage and region (Mpanga versus Nabugabo) on mean embryo size per population with mean DO per site included as a covariate. In a second analysis, the four sample sites where the mean DO was equal to or less than 3.0 mg l-1 were classified as low DO, and the other five populations were classified as high DO (high DO ~ 6.0 mg L-1; mean DO values reported in Reardon and Chapman, 2009). We used an analysis of variance (ANOVA) to test for an effect of DO (high, low) and region on mean embryo size (mass and length) across developmental stages with individual population in-cluded as a random factor (nested within oxygen regime and region). We used the mean mass and length per brood. 1.2 Rearing experiment

Adult live specimens were wild-caught from the hy-poxic Lwamunda Swamp that surrounds Lake Nabu-gabo, Uganda and transferred to our aquatic facility at McGill University. In the swamp, DO levels average 1.85 mg L-1 throughout the year (Reardon and Chapman, 2008). This population exhibits larger gills, higher he-matocrit, and a smaller size at maturity than con-specifics that inhabit well-oxygenated lakes and rivers, traits that may represent an adaptive response to hy-poxic stress (Chapman et al., 2000; Chapman et al., 2002a,b; Chapman et al., 2008; Martinez et al., 2009;

Reardon and Chapman, 2009). Four families of the Lwamunda Swamp population were bred under nor-moxic conditions (Fig. 1). The F1 broods were then split into two groups, and full siblings were raised under high or low DO. One family was raised per aquarium per treatment. We reduced broods randomly to 8 juveniles per aquarium after 3 months. This density allows fish to grow to a size reflective of field populations. The DO in the experimental growth tanks was controlled by Point Four System’s PT4 Monitor by automatically pulsing nitrogen through a diffuser in each aquarium to reduce DO levels. Mean (± SE) DO was 7.98 ± 0.02 mg L-1 in the normoxic aquarium and 1.31 ± 0.00 mg L-1 (n = 537) in the hypoxic aquaria, representative of field DO levels (Reardon and Chapman, 2009). Aquaria were held at 24.9 ± 0.1 ◦C (recorded once daily) and exposed to a 12/12 photoperiod. Underwater filtration was achieved with Hagan Fluval 1+ Underwater pumps with sponge biofiltration. pH, ammonia, nitrate, and conductivity were monitored biweekly. Juvenile fish were fed fry bites daily for the first two weeks post-release from the mouth then switched to Tetramin flakes once per day.

F1 females were observed daily for evidence of brooding (not feeding, dropped jaw, secretive). Ten fe-males produced multiple broods representing three out of the four families grown in each treatment. The fourth family produced no brooders in either DO treatment. The first brood from each mother was collected within the first 24 hr after fertilization; egg metabolic rate was measured individually on 5 eggs per brood. Egg meta-bolic rates were measured using LoligoSystems micro-glass respirometer (volume: 0.7 ml) and AutoResp software using intermittent closed respirometry and a LoliTemp temperature controller to maintain target temperature to ± 0.1 ◦C. Metabolic rates were measured at mean (± SE) temperature of 24.61 ± 0.06 ◦C (±SE; n = 32 eggs; temperature logged once per second through-out trials) in nearly saturated water (8.2 mg L-1). All eggs were held for 1 hr under high DO before respi-rometry measurements. Each egg experienced a 20-minute acclimation period in the respirometry chamber before measurements began. Metabolic rate was calculated based on the average metabolic rate across seven 10-minute measurement periods and was expressed as ugO2 hr-1. Levels of background respira-tion were estimated from controls run in the empty res-pirometer. All eggs were euthanized in MS222 and then preserved in 4% paraformaldehyde buffered with PBS for estimates of egg size (mass and volume). Each egg was placed in a drying oven for 24 hr at 60 °C to


achieve dry mass. Individual egg wet and dry mass were measured to 0.001 mg using a Sartorius Microbalance. Percentage of water was calculated as: (Wet mass – Dry mass) / Wet mass. Individual egg volume was calculated using the equation for a spheroid [V= 4/3*л*(D1/2)* (D2/2)*((D1 + D2)/4)] (Niciu, 1995). D1 and D2 (two perpendicular diameters of each egg) were measured using digital photographs and the imaging software, Motic Images 2000 1.3©. When the second brood of each female was released from the mother’s mouth, fry were collected, euthanized in MS222, and preserved for estimates of juvenile size at release (total length, mass). When the third brood was released from each mother’s mouth, they were split and grown for an additional three months under either low 1.527 ± 0.003 mg l-1 (±SE) or high (7.84 ± 0.06 mg l-1) DO (Fig. 1). Mean (± SE) temperature was 25.98 ◦C ± 0.08 in the low-DO growth tanks and 25.85 ◦C ± 0.09 in the high-DO growth tanks. Three months post-release, juveniles were collected and preserved for estimates of size at time and survival. In-dividual juvenile mass was recorded on a Mettler Bal-

ance, AC-100. Total length was measured using digital photographs and the imaging software, Motic Images 2000 1.3©. Brooding period (from egg fertilization to release from mother’s mouth) was calculated using the second and third broods of each female.

ANOVA was used to test for an effect of the treat-ment under which the F2 young were brooded (devel-opment DO: high or low) on mean egg size (wet mass, dry mass, volume, % water) and mean F2 juvenile size at release (total length, mass) with family and female (nested within family and development DO) included as random factors. Female is defined as the mother of each brood, and family is defined as groups of individuals originating from the same wild-caught parents. ANOVA was also used to test for an effect of treatment on brooding period (time in which the F2 offspring were held in the mouth of the females from fertilization to release) with family included as a random factor. For F2 juveniles grown for 3 months post-release, ANOVA was used to test for an effect of development treatment (high or low DO), post-release growth treatment (high or

Fig. 2 Map of Katonga River system in Uganda, a river that flows both east and west due to the historical rift uplifting (adapted from Reardon and Chapman 2009 and Crispo and Chapman 2010) One sampling (Lake Saka) is indicated here. Inset A illustrates the location of our four sampling sites (asterisks) on or near the Mpanga River that flows into the westward-flowing Katonga River and eventually into Lake George. Inset B illustrates the four sampling sites in the Lake Nabugabo region where the Katonga River meets lowlands and flows into Lake Victoria.


low DO), and their interaction on individual mass and total length, with female and family included as random factors. Similarly, ANOVA was also used to test for an effect of development treatment (high or low DO), post-release growth treatment (high or low DO), and their interaction on brood survival with family included as a random factor. ANCOVA was used to test for an effect of DO treatment and family on individual egg metabolic rate with individual egg mass as a covariate. No egg/juvenile size traits were related to female stan-dard length, although our range in female length was small.

2 Results 2.1 Field survey

In total, 113 brooding females were collected across populations and years. We measured each individual embryo from every brood collected, ranging between 25–37 embryos per female. The interaction between DO and developmental stage was not significant (mass P = 0.752; length P = 0.508), nor between DO and region (mass P = 0.708; length P = 0.710), indicating homoge-neous slopes. There was a positive relationship between mean embryo size and mean DO across developmental stages (ANCOVA for embryo mass: DO F1,19 = 5.258, P = 0.033, stage F3,19 = 2.987, P = 0.057, region F1,19= 3.121, P = 0.093; ANCOVA for embryo length: DO F1,19

= 5.711, P = 0.027, stage F3,19 = 39.556, P < 0.001; re-gion F1,19= 25.121, P < 0.001; Fig. 3). Embryos from low-DO field populations were smaller (mass and length) than embryos from high-DO field populations across all stages of development (ANOVA mass: DO F1,10.5 = 6.668, P = 0.026; stage F3,101 = 16.928, P < 0.001; pop

F6,101 = 1.611, P = 0.152, region F1,9.5 = 0.450, P = 0.519; length: DO F1,12.8 = 5.937, P = 0.030; stage F3,101 = 83.938, P < 0.001; pop F6,101 = 1.122, P = 0.355; region F1,11.6 = 15.214, P = 0.002; Fig. 4). In both analyses, across developmental stages, embryos from sites in the Mpanga region were longer than embryos from sites in the Nabugabo region, but there was no difference in embryo mass between regions. 2.2 Rearing experiment

The (F2) egg size traits of eggs produced by F1 mothers did not differ significantly between DO treat-ments (Table 2, n = 150 eggs). There was a significant interaction between family and DO treatment for egg % water content (Table 2). However, within each DO treatment, there was no difference in egg % water con-tent among families (Sidak post hoc: High DO P = 0.119; Low DO P = 0.098). Metabolic rate was higher in the eggs of females reared under low DO (mean ± SE, 1.00 ± 0.053 µgO2 hr-1) compared to eggs from females reared under high DO (mean ±SE, 0.51 ± 0.054 µgO2 hr-1; n = 32 eggs; ANCOVA: treatment F1,5 = 24.366, P = 0.016). Slopes were homogenous (P = 0.313) between treatment groups. There was no effect of egg mass on egg metabolic rate (P = 0.656).

The total length and wet mass of F2 juveniles at the time of release from the mother’s mouth did not differ between DO treatments (Table 2; n = 82 juveniles). However, brooding period was 16% shorter in broods produced in the low-DO treatment compared to broods produced under high DO (Table 2). Offspring from both DO treatments were at the same macroscopic develop-mental stage (juveniles) when released from the mouth of females.

Fig. 3 Relationship between (a) mean embryo mass, (b) mean embryo length and DO across field populations of Pseudo-crenilabrus multicolor DO data were adapted from Reardon and Chapman (2009).


Fig. 4 Mean length (panel A) and mass (panel B) of Pseudocrenilabrus multicolor embryos of four development stages from high and low-DO field populations in Uganda Means were produced by ANOVA testing for an effect of DO (high, low), controlling for potential region effects across developmental stages with individual population included as a random factor (nested within oxygen regime and region). Error bars represent standard errors. Stars (*) represent a significant difference in embryo mass or length between high and low DO populations within each stage. As reported in the text, within each stage, trends were similar for mean total length between high- and low-DO populations.

Regardless of the DO environment in which they

were brooded, F2 juveniles grown for 3 months under low DO were shorter than full siblings grown under high DO (Fig. 5, Table 3; n = 71 juveniles). There was no significant difference in mass between the growth DO treatments or developmental DO treatments, although the trend was similar to length (Fig. 5, Table 3). Female had a significant effect on embryo mass and length (Ta-ble 3) indicating some variation among females in their offspring size. There was no effect of embryo develop-ment DO, juvenile growth DO, or their interaction on survival (Table 2). After 3 months of post-release juve-nile growth, mean survival was 66.5% (±7.7 SE) under high DO and 77.7% (±7.6 SE) under low DO.

3 Discussion 3.1 Development in the mouth

In field populations of P. multicolor, embryos from low-DO field populations were smaller (mass and length) than embryos from high-DO field populations across all stages of development. A significant region effect was detected, driven by longer embryos in sites of the Mpanga region. However, there was no difference in embryo mass between regions, suggesting a potential difference in embryo shape/morphology between these two regions. Regardless of region, our results support the positive relationship between egg size and DO re-ported in Reardon and Chapman (2009), suggesting that in wild populations, smaller eggs are indeed producing smaller embryos. The observed interpopulational and regional level variation could be genetically determined and/or environmentally induced.

Because of the difficulties of quantifying develop-ment, growth, and survival in wild populations, we used a laboratory rearing experiment to quantify potential environmental influences of DO for one low-DO field population. The results indicated that while in the pro-tection of the mother’s mouth, DO did not affect egg and embryo size; however, size had diverged three months after the juveniles were released from the mother’s mouth. The lack of treatment effect on egg and juvenile size-at-release in the lab suggests that the differences in egg and embryo size across field populations may be genetically-determined. In an earlier study, Reardon and Chapman (2009) reported larger egg size in F1’s of high-DO origin than in F1s of swamp origin supporting a population effect on egg size. The difference in egg volume between DO treatments detected by Reardon and Chapman (2009) was due to a difference in water content of the eggs-a difference we were not able to rep-licate in this study. Experimental studies on the African cichlid Simochromis pleurospilus found that mothers prepare their offspring for the growth environment they experience as juveniles, rather than the environment they experience during adulthood and/or spawning (Ta-borsky, 2005; 2006). If true in P. multicolor, we should have been able to induce a treatment response in this study. Our data suggest that this is not the case. The lack of treatment response could suggest that a mother’s preparation of offspring for their juvenile environment may arise due to long-term genetic selection, rather than environmental response. Future studies that rear off-spring from multiple populations are needed to address these possibilities.


Table 2 Mean trait values for Pseudocrenilabrus multicolor broods mouth brooded by mothers reared under high or low DO and results for ANOVA testing for an effect of embryo developmental DO environment on size traits with family and female (nested within family and developmental DO environment) included as random factors

Egg/Juvenile at release Trait Development DO Mean Trait Value (± SE) Factor F df P

DO Treatment 3.844 1,1.8 0.203 High DO 5.73 (± 0.03)

Family 2.111 2,1.6 0.356

Female 24.272 7,136 <0.001*Egg wet mass (mg)

Low DO 5.29 (± 0.05) DO*Family 0.852 2, 6.9 0.467

High DO 2.81 (± 0.02) DO Treatment 37.344 1,1.2 0.081

Family 44.575 2, 0.5 0.251

Low DO 2.59 (± 0.03) Female 16.549 7, 135 <0.001*Egg dry mass (mg)

DO*Family 0.192 2, 6.9 0.830

High DO 0.51 (± 0.003) DO Treatment 0.024 1, 1.9 0.891

Family 0.202 2, 1.9 0.832

Low DO 0.51 (± 0.002) Female 3.342 7, 135 0.003* % Water

DO*Family 22.115 2, 6.6 <0.001*


Family 0.088 2, 1.7 0.920

Low DO 6.24 (± 0.09) Female 10.929 7, 136 <0.001*Volume (μL)

DO*Family 1.212 2, 6.9 0.354

High DO 10.42 (± 0.12) DO Treatment 1,1.9 0.126 0.758

Family 2,1.5 5.050 0.219

Low DO 9.97 (± 0.11) Female 3,63 44.483 <0.001*Juvenile at release wet mass (mg)

DO*Family 2,3 0.303 0.759


Family 1.276 2, 1.9 0.443

Low DO 8.77 (± 0.08) Female 8.491 4, 71 <0.001*

Juvenile at release total length (mm)

DO*Family 2,4.1 0.880 0.481

High DO 19.22 (± 1.22) DO Treatment 10.904 1, 2.9 0.047*

Family 4.333 2, 2 0.187 Brooding period

Low DO 16.22 (± 1.28) DO*Family 0.222 2, 7 0.801

The stars (*) represent significant P values.

Despite the lack of difference in egg or juvenile size-at-release between treatments, the brooding period data suggest the possibility that development may be faster under low DO. Embryos brooded by mothers that were reared under low DO left the mouth 16% earlier than embryos brooded by mothers reared under high DO, but they were at the same stage of development (yolk sac fully absorbed) when released from the mouth. A similar finding was reported in wild-caught P. multi-color acclimated to high and low DO (Reardon and Chapman, 2010). Although future studies are clearly necessary to examine development at a finer scale (e.g., histological studies), to our knowledge, this is

among the first evidence for faster development of fish in response to development under hypoxia and is in contrast to other reports of delayed development under hypoxia in fishes (Garside, 1959; Hamor and Garside, 1976; Rombough, 1988; reviewed in Wu, 2009; Mueller et al., 2011). However, our study of P. multicolor is also the first to explore these issues in a bearer fish species that has lived and evolved under low-DO for many gen-erations in the wild; it may be that parental care plays a role in developmental response to hypoxia. In their study of the costs of mouth brooding in P. multicolor, Reardon and Chapman (2010) found that the standard metabolic rates were ~48% higher in brooding females


Fig. 5 Mean mass and length of F2 Pseudocrenilabrus multicolor juveniles grown for 3 months post-release from the mouth under either high or low DO The star (*) represent a significant difference between high- and low-DO post-release juvenile growth environments within each size trait. Error bars represent standard errors.

Table 3 Results for ANOVA testing for an effect of DO environment (high or low DO) during embryo development in the mother’s mouth (development DO), three months of growth after release from the mother’s mouth (growth DO), and their interaction on total length and mass for F2 P. multicolor with family and female (nested within family and developmental DO) included as random factors

Size Trait Factor df F P

Total length (mm) Development DO 1, 3.8 0.101 0.768

Growth DO 1, 61 4.129 0.047*

Development*Growth DO 1, 61 0.722 0.399

Family 2, 4.4 0.377 0.706

Female 4, 61 2.575 0.046*

Mass (mg) Development DO 1, 3.8 0.121 0.746

Growth DO 1, 61 2.311 0.134


Family 2, 4.3 0.129 0.883

Female 4, 61 3.232 0.018*

Survival Development DO 1, 10 0.862 0.311

Growth DO 1, 10 1.139 0.375


Family 2, 10 9.954 0.004*

The stars (*) represent significant P values. compared to post-brooding females, suggesting a high energetic cost to the female during brooding. It is possi-ble that P. multicolor females compensate for this cost

with a reduced brooding period (Reardon and Chapman, 2010). There may also be an advantage to offspring of moving quickly through this developmental stage in a


hypoxic environment. Clearly more work is necessary to tease apart whether the reduced brooding period/embryo development time is a parental, offspring, or mutual adaption to hypoxia. The mechanisms underlying the shorter brooding period also remain unknown. In a study exploring the developmental plasticity of the car-diovascular system in zebrafish embryos grown under high or low DO, Pelster (2003) found that four days post-fertilization, embryos developing under low DO had enhanced blood vessel formation and increased car-diac output compared to those under high DO. Given the positive link between cardiac ouput and metabolism in the literature, Pelster’s results are supportive of our increased metabolic rate in low DO eggs and the poten-tial link to faster development rate. A study on brine shrimp Artemia franciscana also reported faster deve-lopment in shrimp cultured under chronic hypoxia rela-tive to high DO due to earlier respiratory regulation in the hypoxia-cultured individuals. The improved respi-ratory regulation in the hypoxia-reared brine shrimp took place before development of the heart and gills suggesting that observed shifts in haemoglobin concen-tration may be the mechanism driving the faster deve-lopment under chronic hypoxia (Spicer and El-Gamal, 1999). If respiratory regulation is conserved across taxa, perhaps a similar shift in haemoglobin concentration is occurring in P. multicolor. In adult P. multicolor, Martínez et al. (2009) showed that fish from the same population as this study had elevated haematocrit and LDH activities when compared to high-DO populations of the same species. More work is needed to explore the possibilities of the hypoxia tolerance mechanisms acting in developing embryos. 3.2 Development outside of the mouth

F2s brooded by low-DO reared F1 mothers exhibited a divergence in their growth measured 3 months after they were released; juveniles grown under high DO were, on average, ~10% longer than juveniles grown under low DO. These findings are supportive of the pre-dictions for smaller size under low DO and consistent with the positive relationship with embryo size and DO across field populations. Mass did not diverge signifi-cantly between juvenile growth environments; but the trend was in the same direction as length. Although we were unable to measure juveniles in a comparable post-release stage in field populations, the trends for egg and embryo size support a smaller juvenile size in hy-poxic field populations. Low-DO field populations had smaller eggs and embryos relative to high-DO popula-tions whereas, in the lab rearing experiment, egg and

embryo size was the same between high- and low-DO development environments. Thus, differences in post- release juvenile size in the field may be more extreme than differences reported in this lab rearing study. To explore this possibility, juvenile size will need to be quantified across field populations after growth outside the mouth of the mother. Trade-offs between high- and low-DO environments may contribute to variation in embryo and juvenile size among field populations. Large juveniles from high-DO habitats may suffer res-piratory distress if they moved into hypoxic waters, while smaller sized juveniles from low DO may perform sub-optimally if they moved to high-DO waters associ-ated with other ecological factors. For example, there is a growing body of literature suggesting that predation pressure of fish from piscine predators is reduced under low DO (Robb and Abrahams, 2003, Chapman and McKenzie, 2009). For the Lwamunda Swamp popula-tion, predation pressure from piscine predators is likely to be very low (Chapman, Chapman and Chandler, 1996a; Chapman et al., 1996b; 2002a). The main preda-tion pressure in this population is aerially from king-fishers (Randle and Chapman, 2004), which do not tar-get the juveniles, thus reducing negative consequences of small size under low DO. P. multicolor from high-DO river and lake sites live in sympatry with other cichlids that consume juveniles (Binning and Chapman, 2008), and thus bigger should be better in the high-DO habitats (Reznick, 1991) and smaller better in low-DO habitats (Reardon and Thibert-Plante, 2010). It would be inter-esting to explore these questions in a high-DO popula-tion of P. multicolor and in other bearer species of fish to see if the P. multicolor’s response to development and growth under low DO is widespread across populations and in other bearer species from a diversity of lineages.

It is important to note that some studies report slower growth and/or smaller size under hypoxia as evidence of negative fitness effects (e.g. Wu, 2009; Mueller et al., 2011) i.e. there is no advantage to being small under hypoxia, rather small size is a limitation of the growth environment. Future studies that quantify metabolic rates of juveniles from both high- and low-DO popula-tions under high and low DO may be useful in inform-ing our understanding of factors driving small size un-der hypoxia in eggs and embryos. Our findings support the predictions in the hypoxia and development litera-ture, providing evidence for smaller body size across embryo stages under hypoxia in both the laboratory and field. In contrast to the literature, we found no evidence for slower development or lower survival under hypoxia.


Many of the studies reporting reduced fitness under hy-poxia were conducted on species with no post-spawning parental care. Yet, a high proportion of parental care species occur in low DO habitats suggesting that paren-tal care may offset negative impacts of hypoxia on their offspring (Chapman and McKenzie, 2009). This work provides insights into how bearer fishes may respond to increases in global hypoxia and suggest that compared to species with no parental care, bearer fishes may be more able to adapt to long-term hypoxia in their envi-ronment.

Acknowledgements We thank Colin Chapman, Patrick Omeja and the Kibale/Lake Nabugabo field teams for assis-tance with the field collections; Lubov Grigoryeva for assis-tance in embryo size data collection. Funding for this study was provided from the McGill International Doctoral Award (ER), the Wildlife Conservation Society (LC), the Natural Sciences and Engineering Research Council of Canada (Dis-covery Grant, LC), McGill University, and the Canada Re-search Chair program (LC). Permission to conduct research was acquired from the Uganda National Council for Science and Technology (UNCST EC 630) and the McGill University Animal Care Committee (Protocol #5029). We thank members of the L. Chapman lab for comments on previous versions of this manuscript.

References

Balon E, 1986. Types of feeding in the ontogeny of fishes and the life-history model. Environmental Biology Fishes 16: 11–24.

Binning SA, Chapman LJ. 2008. Feeding ecology and diet overlap in riverine cichlids from western Uganda. Verhandlungen In-ternationale Vereinigung Limnologie 30: 283–286.

Bradford DF, 1990. Incubation and rate of embryonic develop-ment in amphibians: The influence of ovum size, temperature, and reproductive mode. Physiological Zoology 63: 1157–1180.

Chapman LJ, McKenzie D, 2009. Behavioural responses and ecological consequences. In: Richards JG, Farrell AP, Brauner CJ ed. Hypoxia in Fishes. San Diego: Elsevier, 26–77.

Chapman LJ, Chapman CA, Chandler M. 1996a. Wetland ecotones as refugia for endangered fishes. Biological Conser-vation 78: 263–270.

Chapman LJ, Chapman CA, Ogutu-Ohwayo R, Chandler M, Kaufman L et al., 1996b. Refugia for endangered fishes from an introduced predator in Lake Nabugabo, Uganda. Conserva-tion Biology 10: 554–561.

Chapman LJ, Galis F, Shinn J, 2000. Phenotypic plasticity and the possible role of genetic assimilation: Hypoxia-induced trade- offs in the morphological traits of an African cichlid. Ecology Letters 3: 387–393.

Chapman LJ, Chapman CA, Nordlie FG, Rosenberger AE, 2002a. Physiological refugia: Swamps, hypoxia tolerance, and main-

tenance of fish biodiversity in the Lake Victoria region. Com-parative Biochemistry and Physiology A 133: 421–437.

Chapman LJ, Nordlie FG, Seifert A. 2002b. Respiratory oxygen consumption among groups of Pseudocrenilabrus multicolor victoriae subjected to different oxygen concentrations during development. Journal of Fish Biology 61: 242–251.

Chapman LJ, Albert J, Galis F. 2008. Developmental plasticity, genetic differentiation, and hypoxia-induced trade-offs in an African cichlid fish. Open Evolution Journal 2: 75–88.

Crispo E, Chapman LJ, 2010. Geographic variation in phenotypic plasticity in response to dissolved oxygen in an African cichlid fish. Journal of Evolutionary Biology 23: 2091–2103.

Diaz RJ, Breitburg DL, 2009. The hypoxic environment. In: Richards JG, Farrell AP, Brauner CJ ed. Hypoxia in Fishes. San Diego: Elsevier, 1–23.

Einum S, Hendry AP, Fleming IA, 2002. Egg size evolution in aquatic environments: Does oxygen availability constrain size. Proceedings of the Royal Society of London B 269: 2325– 2330.

Fryer G, Iles TD, 1972. Breeding habits. In: Fryer G, Iles TD ed. The Cichlid Fishes of the Great Lakes of Africa: Their Biology and Evolution. Edinburgh: Oliver and Boyd, 105–172.

Garside ET, 1959. Some effects of oxygen in relation to tempera-ture on the development of lake trout embryos. Canadian Journal of Zoology 37: 689–698.

Hamor T, Garsdie ET, 1976. Developmental rates of embryos of Atlantic salmon Salmo salar L. in response to various levels of temperature, dissolved oxygen, and water exchange. Canadian Journal of Zoology 54: 1912–1917.

Heath DD, Blouw DM, 1998. Are maternal effects in fishes adap-tive or merely physiological side effects? In: Mousseau TA, Fox CW ed. Maternal Effects as Adaptations. Oxford: Oxford University Press, 178–201.

Hellweg M, 2008. The dwarf Victoria mouth brooder: A new look at a classic fish. Tropical Fish Hobbyist Sept: 90–93.

Hendry AP, Day T, Cooper AB, 2001. Optimal size and number of propagules: Allowance for discrete stages and effects of ma-ternal size on reproductive output and offspring fitness. American Naturalist 106: 387–407.

Kolm N, Ahnesjo I, 2005. Do egg size and parental care coevolve in fishes? Journal of Fish Biology 66: 1499–1515.

Krogh A, 1959. The Comparative Physiology of Respiratory Mechanisms. Philadelphia: University of Pennsylvania Press.

Lindsey CC, 1978. Form, function and locomotory habits in fish. In: Hoar WS, Randall DJ ed. Fish Physiology, vol. VII. Lon-don: Academic Press, 1–100.

Marshall DJ, Bolton TF, 2007. Effects of egg size on the deve-lopment time of non-feeding larvae. Biological Bulletin 212: 6–11.

Marshall DJ, Keough MJ, 2008. The evolutionary ecology of offspring size in marine invertebrates. Advances in Marine Bi-ology 53: 1–50.


Martinez ML, Chapman LJ, Rees BB, 2009. Population variation in hypoxic responses of the cichlid Pseudocrenilabrus multi-color victoriae. Canadian Journal of Zoology 87:188–194.

Moyle PB, Cech JJJ, 1996. Respiration. In: Moyle PB, Cech JJJ ed. Fishes: An Introduction to Ichthyology. 3rd edn. Upper Saddle River: Prentice Hall, 37–49.

Mueller CA, Seymour RS, Joss JMP, 2011. Effects of environ-mental oxygen on development and respiration of Australian lungfish Neoceratodus forsteri embryos. Journal of Compara-tive Physiology B. DOI 10.1007/s00360-011-0573-3

Niciu E, 1995. An analysis of maternal investment in the mum-micog (Fundulus heteroclitus; Pisces: Cyprindontidae). Master Thesis, University of Florida, 55.

Noakes DLG, 1991. Ontogeny of behaviour in cichlids. In: Keenleyside MHA ed. Cichlid Fishes: Behaviour, Ecology and Evolution. London: Chapman and Hall, 210–211.

Pauly D, 1981. The relationships between gill surface area and growth performance in fish: A generalization of von Berta-lanffy's theory of growth. Berichte der Deutschen Wissench-aftlichen Kommission für Meeresforschung 28: 251–282.

Pauly D, Pullin RSV, 1988. Hatching time in spherical, pelagic, marine fish eggs in response to temperature and egg size. En-vironmental Biology of Fishes 22(4): 261–271.

Pelster B, 2003. Developmental plasticity in the cardiovascular system of fish, with special reference to the zebrafish. Com-parative Biochemistry and Physiology Part A 133: 547–553.

Pollock MS, Clarke LMJ, Dube MG, 2007. The effects of hypoxia on fishes: From ecological relevance to physiological effects. Environmental Reviews 15: 1–14.

Przeslawski R, Webb AR, 2009. Natural variation in larval size and developmental rate of the northern quahog Mercenaria mercenaria and associated effects on larval and juvenile fitness. Journal of Shellfish Research 28: 505–510.

Randle AR, Chapman LJ, 2004. Habitat use by the air-breathing fish Ctenopoma muriei: Implications for costs of breathing. Ecology of Freshwater Fish 13: 37–45.

Reardon EE, Chapman LJ, 2008. Reproductive seasonality in a swamp-locked African cichlid. Ecology of Freshwater Fish 17: 20–29.

Reardon EE, Chapman LJ, 2009. Hypoxia and life-history traits in a eurytopic African cichlid. Journal of Fish Biology 75: 1795–1815.

Reardon EE, Chapman LJ, 2010. Hypoxia and energetic of mouth

brooding: Is parental care a costly affair? Comparative Bio-chemistry and Physiology A 156: 400–406.

Reardon EE, Thibert-Plante X, 2010. Optimal offspring size in-fluenced by the interaction between dissolved oxygen and predation pressure. Evolutionary Ecology Research 12: 377–387.

Reznick D, 1991. Maternal effects in fish life histories. In: Dudley E ed. Proceedings of the IVth International Congress of Sys-tematic and Evolutionary Biology. Portland: Dioscorides Press, 780–793.

Rius M, Turon X, Dias GM, Marshall DJ, 2010. Propagule size effects across multiple life-history stages in a marine inverte-brate. Functional Ecology 24: 685–693.

Robb T, Abrahams MV, 2003. Variation in tolerance to hypoxia in a predator and prey species: An ecological advantage of being small. Journal of Fish Biology 62: 1067–1081.

Rombough PJ, 1988. Growth, aerobic metabolism, and dissolved oxygen requirements of embryos and alevins of steelhead Salmo gairdneri. Canadian Journal of Zoology 66: 651–660.

Rombough PJ, 2007. Oxygen as a constraining factor in egg size evolution of salmonids. Canadian Journal of Fisheries and Aquatic Sciences 64: 692–699.

Sargent RC, Taylor PD, Gross MR, 1987. Parental care and the evolution of egg size in fishes. American Naturalist 129: 32–46.

Spicer JI, El-Gamal MM, 1999. Hypoxia accelerates the deve-lopment of respiratory regulation in brine shrimp - but at a cost. Journal of Experimental Biology 202: 3637–3646.

Taborsky B, 2005. The influence of juvenile and adult environ-ments on life-history trajectories. Proceedings of the Royal Society of London B 273: 741–750.

Taborsky B, 2006. Mothers determine offspring size in response to own juvenile growth conditions. Biology Letters 2: 225–228.

Welcomme RL, 1969. The biology and ecology of the fishes of a small tropical stream. Journal of Zoological 158: 485–529.

Wickler W, 1962. Zur Stammesgeschichte funktionell korrelierter organ-und verhaltensemerkmale: ei Attrappen und maulbrüten bei afrikanischen cichliden. Zeitschrift fur Tierpsychologie 19: 129–164.

Wootton RJ, 1990. Ecology of Teleost Fishes. London: Chapman and Hall.

Wu RSS, 2009. Effects on fish reproduction and development. In: Richards JG, Farrell AP, Brauner CJ ed. Hypoxia in Fishes. San Diego: Elsevier, 79–141.

Fish embryo and juvenile size under hypoxia in the mouth...

Documents

Transcript of Fish embryo and juvenile size under hypoxia in the mouth...